Lido yields 3%. Ondo yields 4-5%. Ethena yields double digits. Comparing those numbers is comparing four different risks. The question is not which is biggest. It is which risk you want on your balance sheet, and how much you should be paid to hold it.
APY is the most misread number in crypto. It is not a return; it is a bill. Lido, Ondo, Ethena, and Pendle are four different ways of packaging risk, and the APY on each is the price the market quotes for warehousing that risk. Rank them by yield and you rank them by exactly the thing you should be sizing down.
Lido's yield comes from securing Ethereum. Ondo's yield comes from the U.S. Treasury. Ethena's yield comes from derivative market structure. Pendle's yield comes from whatever asset sits underneath it. Three of these are yield sources. Pendle is a yield marketplace. The asymmetry matters and we will come back to it.
The 9-point spread between Lido's 3% and a recent Ethena print in the low teens is not a gift. It is the spread for warehousing perpetual-funding regime risk that Lido holders do not carry. The 1-point spread between Lido and Ondo is not free either. It is the spread between staking-execution risk and sovereign credit, paid in opposite directions depending on your view of the dollar.
That sounds provocative. It is actually the most boring observation in finance. Here is the credit ladder, in any era, in any asset class:
The number on top is the toll you pay for sitting on what is underneath. Crypto did not invent this principle. It is rediscovering it.
You hand Lido ETH. Lido delegates validator operation to a permissioned node-operator set. You receive stETH that grows balance-style as those validators earn issuance, priority tips, and MEV, less Lido's fee. The yield is stable across cycles because it is a function of network economics, not market structure. The principal risk is slashing or validator-set failure, both slow-moving and unrealized to date.
Ondo holds short-duration Treasury exposure (directly for institutional OUSG, via a BlackRock fund for retail USDY) and tokenizes the claim. The yield is the T-bill rate. The risk is sovereign duration and inflation, plus wrapper-mechanics (custody chain, redemption rails, regulatory status). Ondo's tokens today are used mostly in permissioned or whitelisted contexts; broad DeFi-collateral acceptance is limited but not zero.
Ethena holds spot ETH (and other assets) and shorts the equivalent perpetual on centralized exchanges. The hedge nets price exposure to roughly zero; the position collects the perpetual funding rate. The yield is a function of how many traders want to be long perps with leverage: high single digits in calm markets, 20%+ in stretched-long regimes, compressed in neutral periods, and negative in bear-leaning regimes (when the position pays out instead of taking in). The risks are venue concentration (custody on the exchanges where the shorts live), basis blowout, and the regime dependency itself.
Pendle does not generate yield. It separates a yield-bearing asset into a principal token (PT) you can buy at a discount to face value and a yield token (YT) that captures all the yield for a fixed duration. In a market where every yield has a price, Pendle is the venue where that price is set. Useful when you have a view on where the underlying yield is going; rarely the right hold without one.
The yield gap does not only price risk. It also prices reuse. This part is genuinely under-discussed.
stETH sits in dozens of major DeFi venues as accepted collateral. A sophisticated user deposits stETH, borrows stables against it at 70-90% LTV, and redeploys those stables, all while the underlying stETH keeps earning its staking yield. A unit of stETH does not yield 3%. It yields 3% plus optionality on every venue that accepts it. The market is paying for that optionality when it accepts a lower headline number versus a less-composable competitor.
sUSDe is accepted in fewer venues. Ondo's tokens are usable mostly in permissioned contexts. Both compensate holders with higher headline yield to attract capital that would otherwise prefer a more reusable instrument. The differential is self-reinforcing: composable assets attract capital, deepen liquidity, earn more venue acceptance; the less-composable have to keep paying up to compete. This is not an arbitrage waiting to close. It is the structural pricing of two different products that happen to share the label "yield."
These instruments are not competing products. They serve different portfolio functions and the right allocation depends on what the allocator is trying to do.
| Instrument | Portfolio Function | Sized Like | You Own It When |
|---|---|---|---|
| Lido (stETH) | Core crypto-yield exposure | Long-duration position | You want ETH price exposure with a base layer of staking yield, and you want collateral that works in DeFi |
| Ondo (OUSG / USDY) | Defensive USD-yield exposure | Cash-equivalent position | You want T-bill rate inside the onchain wrapper without leaving custody, and you accept reduced DeFi composability for sovereign credit |
| Ethena (sUSDe) | Aggressive carry-trade exposure | Hedge-fund position | You have conviction that funding stays structurally positive across the holding period, and you can monitor venue concentration + basis-stress conditions |
| Pendle (PT / YT) | Tactical yield-curve positioning | Trade, not a hold | You have a directional view on whether the underlying yield rises or falls, or you want to lock in a known fixed yield for a defined horizon |
Most allocators treat these as a single bucket called "yield" and rebalance toward whatever is highest at the moment. The result is a portfolio whose risk profile shifts every time the APY ranking shifts, usually toward the asset with the most fragile yield architecture at exactly the wrong moment. The discipline is to hold each instrument for the function it serves, not for the rank-ordered yield.
The protocols change every cycle. The mechanism does not:
Investors rank APYs across the available options.
Capital flows toward the highest number on the board.
The conditions producing that yield change.
The headline number collapses. Often quickly.
Investors discover they were never buying yield. They were buying risk.
The clearest example crypto produced is stETH in June 2022. For two years, stETH had traded at roughly 1:1 with ETH because every major DeFi venue accepted it as collateral and the staking yield was reliable. The composability was the product; the 4% APY was a bonus on top.
Then Luna failed. Then 3AC failed. Then Celsius needed to exit a large stETH position to meet redemptions. stETH did not break. The validators kept earning, the smart contract kept working, the underlying ETH was still there. What changed was that suddenly every leveraged holder wanted out at the same time and the composability route required unwinding through Curve's stETH/ETH pool, which was not deep enough to absorb forced sellers without slippage. The pool drained, the price detached, stETH traded down to roughly 0.94 ETH.
Investors who said stETH "broke" misread what happened. stETH did not break. The yield was always compensation for two things: validator-execution risk (rare, slow-moving) and exit-liquidity risk (rare, fast-moving). For two years only the first risk seemed to matter, because the second never showed up. When it did, the price moved instantly to reflect what had always been priced in. The yield did not evaporate. The risk just stopped being hypothetical.
That is what every yield-chasing cycle looks like up close. The investor who treats Lido, Ondo, Ethena, and Pendle as four versions of the same product ranked by APY is running the same loop the stETH chasers ran in 2022, just with different protocols. The investor who reads the yield as a risk signal asks the only question worth asking: what am I being paid to hold?
Lido, Ondo, Ethena, and Pendle package four different risks: ETH security, sovereign duration, derivative regime, and yield-curve. The APY is the bill the market sends you for warehousing each one. Higher APY does not mean higher return. It means the market wants more compensation to hold what is underneath.
So APY is the wrong question. The right question is what you are being paid to hold, whether you want to hold it, and whether the price is fair. The cleanest way to fail at yield investing is to optimize for the number on top. The cleanest way to do well is to read it as a disclosure and decide whether the underlying belongs on your balance sheet. Everything else is APY screenshots.